Měření a zpracování dat na MR v laboratorních podmínkách

Abstract

This bachelor thesis deals with measurement and data processing on an MRI practice system in laboratory conditions. In the theoretical part, I outlined the basic tenets of magnetic resonance which are the key to the whole domain of this topic. The thesis also includes a recherche for various methods of image segmentation originating in MR. The practical part of the thesis is focused on working with the MRI practice system, measuring signals and processing images created in this prototype. The aim of the work was segmenting objects from MR in the Matlab interface, generate an artificial noise on created images and then reduce noise by subsequent filtration. For segmentation, I have chosen the method of active contours based on knowledge from the recherche part of the thesis. I have loaded the images with artificial noise – salt and pepper, Gaussian and Poisson type. To determine the similarity of native and artificial noisy images, I used the method of structural similarity of SSIM and the mean square error of MSE. The thesis includes a laboratory task which has a partially educational character and it serves to acquire knowledge and expand comprehension of MRI